ASRock.com Homepage
Forum Home Forum Home > Technical Support > Intel Motherboards
  New Posts New Posts RSS Feed - Asrock X99X Killer - SATA Ports
  FAQ FAQ  Forum Search Search  Events   Register Register  Login Login

Asrock X99X Killer - SATA Ports

 Post Reply Post Reply Page  <12
Author
Message
parsec View Drop Down
Moderator Group
Moderator Group
Avatar

Joined: 04 May 2015
Location: USA
Status: Offline
Points: 4996
Post Options Post Options   Thanks (0) Thanks(0)   Quote parsec Quote  Post ReplyReply Direct Link To This Post Posted: 24 Jul 2015 at 3:32pm
Before I show my testing results, which we can decide if it has any significance, I noticed something in an earlier post of yours. That is:

"All drive SATA connections are from the motherboard ports to their corresponding hotswap bay connectors...".

Hot swap bay connectors set off a red flag, please explain what the hotswap bays are. Part of a PC case? An add on piece of hardware? Hotswap bays are known trouble makers. Do you know if they are "active", meaning powered due to their own circuitry, or purely passive, unpowered? I'm aware of hot swap bays that won't allow some drives to operate at all.

Since the Z87 chipset, Intel has adopted some of the SATA 3.1 protocols, which leaves some drives behind, as their controllers are not compatible with those protocols. This includes some SSDs with early SSD controllers. That may also affect some of your drives.

The closest SSD I have to DevillEars' Samsung is my Samsung 840 EVO 250GB SSD. This is the 840 Evo on the SSATA_0 port:



That is a normal result for an 840 EVO, with the CPU idling, so a non-tweaked benchmark result.

Since the result seemed normal, but I was only testing one drive, I ran three instances of AS SSD at the same time, two for drives on the SATA_0 - 5 ports, and the 840 EVO still on the SSATA_0 port. The is the result for the 840 EVO:



Yes, the results are slightly better. That was caused by the load of three instances of AS SSD taking at least one core of the CPU out of its idle, low clock speed, as well as the core not being in a C State power saving state. This benchmark and SATA IO in general is not CPU intensive, at least for high performance processors.

Just for comparison, the SanDisk OS SSD I posted the benchmark for previously took the greatest performance hit, but that was due to it serving as the OS drive with more work to do:



I'm really wondering what your configuration of drives and other hardware really is, which is important to know and may be related to your issues.

The following is from the Intel® C610 Series Chipset and Intel® X99 Chipset Platform Controller Hub (PCH) datasheet:

Serial ATA (SATA) Controller

The PCH has three integrated SATA host controllers that support independent DMA operation on up to total ten ports and supports data transfer rates of up to 6.0 Gb/s on all ports. The SATA controller contains two modes of operation?”a legacy mode using I/O space, and an AHCI mode using memory space. Software that uses legacy mode will not have AHCI capabilities. The PCH supports the Serial ATA Specification, Revision 3.0. The PCH also supports several optional sections of the Serial ATA II: Extensions to Serial ATA 1.0 Specification, Revision 1.0 (AHCI support is required for some elements).

That does not seem to be an on-die SATA hub, or a kludge.

Link to this datasheet: http://www.intel.com/content/www/us/en/chipsets/x99-chipset-pch-datasheet.html

IF it turns out the SATA ports on the X99 chipset are the limiting factor, a true workstation board with the C610 chipset (enterprise version of the X99) may be a better choice for your requirements.
Back to Top
DevillEars View Drop Down
Newbie
Newbie


Joined: 09 Jul 2015
Location: South Africa
Status: Offline
Points: 52
Post Options Post Options   Thanks (0) Thanks(0)   Quote DevillEars Quote  Post ReplyReply Direct Link To This Post Posted: 24 Jul 2015 at 4:54pm
Hi Parsec,
 
Many thanks for your efforts to-date - they are much appreciated.
 
With respect to your question regarding the hotswap bays, these are an integral element of the chassis in use - a Lian-Li PC-Z70 full tower:
 
 
There are three modules, each offering three bays. Each module has three power connectors and three SATA signal connectors - one set per drive bay - and these are all "pass-through" connections.
 
My other system is housed in the "baby brother" of this chassis - Lian-Li's PC-Z60 mid-tower - which has two modules of three bays each and works fine in conjunction with an Asrock Fatal1ty Z77 Professional board and six x 6TB drives with an OCZ Vertex 3 256GB SSD as OS drive.
 
 
The only physical differences between the two setups are the motherboards and the number of modules/drives (Both boards share the same hotswap module design - 2 modules in older system and 3 modules in new one). Connection differences are the most obvious as the older system has the following connection layout:
 
SATA_0: SSD (Primary boot device)
SATA_1: DVD R/W (Secondary boot device)
SATA_2: HDD (6TB Seagate Enterprise)
SATA_3: HDD (6TB Seagate Enterprise)
SATA_4: HDD (6TB Seagate Enterprise)
SATA_5: HDD (750GB Seagate 2.5" 7200rpm Office Data drive)
S_SATA_0: HDD (6TB WD Green)
S_SATA_1: HDD (6TB WD Green)
S_SATA_2: HDD (6TB WD Green)
S_SATA_3: BrD R/W
 
With this connection layout (optimised so that faster drives share the faster Z77 ports and the slower drives share the slower ASMedia ports), I get very satisfactory data transfer rates that reflect the combination of drive speeds and port speeds. Copying from SATA_2 to SATA_3 yields copy rates (ex TeraCopy) of in excess of 200MB/sec, while copying from SATA_2 to S_SATA_0 yields speeds of up to 170MB/sec (both speeds being close to the maximum attainable rate per drive type).  Even USB 3.0 copies show benefits - copying to a 3TB ST3000DM001 (in Zalman USB 3.0 dock) from SATA_2 was hitting 170MB/sec and achieving an overall rate across 2TB of 140MB/sec.
 
So, given that the "hotswap concern" is a common element across two configurations and that one works fine while the other "is on a go-slow strike", I would rule out the hotswap as a contributing factor.
 
The new system was built on Monday/Tuesday this past week with no HDDs available for testing - just the Samsung SSD. The port connection layout is as follows:
 
SATA_0: Wired but no drive
SATA_1: Wired but no drive
SATA_2: DVD/RW
SATA_3: Wired but no drive
SATA_4: Wired but no drive
SATA_5: Wired but no drive
S_SATA_0: Wired but no drive
S_SATA_1: SSD (850PRO)
S_SATA_2: Seagate ST3000DM001
S_SATA_3: Seagate ST3000DM001
 
plus for performance evaluation:
 
Marvell_0: Seagate ST3000DM001
Marvell_1: Seagate ST3000DM001
 
This situation is temporary as the plan is to migrate the 6TB drives from old system to new and to have the drive:port allocations mirror the logic used in older system drive:port allocations. Until, however, this current "go-slow situation" is resolved, I will fill the new system with 3TB drives to be filched from external enclosures.
 
The folks who built the new rig will be back on Monday and will re-configure the port:bay wiring connections as per one of my earlier posts and then I'll be in a better position to assess the situation.
 
The overall configuration of the new rig is as follows:
 
Chassis: Lian-Li PC-Z70
PSU: Corsair AX860
Mobo: Asrock Fatal1ty X99X Killer
CPU: Intel Core i7-5820K (stock clock)
RAM: 4 x 8GB Corsair Vengeance LPX DDR4-2133
GPU: EVGA GTX970 Superclocked version
Cooler: Corsair H80i GT
SSD: Samsung 850PRO
HDD: 4 x Seagate ST3000DM001 (to be upped to a total of 9 drives)
DVD: Asus DW-A24R DVD writer
BRD: LG BluRay Writer (BW-L16R)
SATA: Sunix S2600 2-port PCIe 2.0 1x SATA 6G card
O/S: Windows 8.1 (patched to current level)
 
Plus externals:
 
Microsoft Wireless Desktop 3000 (Kbd & Rodent)
LG IPS236V 23" FHD IPS monitor
 
I hope that is enough info?
 
PS: I take your point regarding "not a kludge" - my bad!
 
Regards,
 
Dave


Edited by DevillEars - 24 Jul 2015 at 4:57pm
If music be the love of food, eat on
Back to Top
DevillEars View Drop Down
Newbie
Newbie


Joined: 09 Jul 2015
Location: South Africa
Status: Offline
Points: 52
Post Options Post Options   Thanks (0) Thanks(0)   Quote DevillEars Quote  Post ReplyReply Direct Link To This Post Posted: 01 Aug 2015 at 11:09pm

SATA performance "cause" determined and issue resolved!

 
Firstly, the drive:port allocation was re-done by re-connecting all drive:port cables so that the connections were as follows (my thanks to my dealer/neighbour who shall remain nameless Wink ):
 
SATA_0 : Samsung 850PRO (System/OS Drive)
SATA_1 : Seagate ST3000DM001 (Data)
SATA_2 : Seagate ST3000DM001 (Data)
SATA_3 : Seagate ST3000DM001 (Data)
S_SATA_1 : Seagate ST3000DM001 (Data)
USB 3.0: Sandisk Extreme 64GB flash drive
 
(Other ports rewired to appropriate empty hotswap bays)
 
Ran Teracopy permutations after re-connections above and all ran at between 60 and 80 MB/sec (from SATA_0 to SATA_2, from USB 3.0 to SATA_0, from SATA_2 to S_SATA_1 etc.)
 
Data being copied comprised 31 x .mkv files totalling 54.1 GB.
 
This seemed suspicious - why should performance be consistently low across connections of different bandwidth capability?).
 
Downloaded Crystal Disk Mark and ran some tests on SATA_0 (SSD), SATA_2 and S_SATA_1 (7200rpm HDDs) with the following results for sequential read:
 
SATA_0: 520 MB/sec
SATA_2: 180 MB/sec
S_SATA_1 : 175 MB/sec
 
This performance seemed better aligned to the bandwidth potential of the drives, so decided to uninstall Teracopy and run copy tests using Win 8.1 File Copy instead:
 
SATA-0 to SATA_2: 183 MB/sec
SATA-0 to S_SATA_1: 179 MB/sec
USB 3.0 to SATA_0: 235 MB/sec
SATA_2 to SATA_1: 165 MB/sec
 
This spread of performance mirrored the pattern from Crystal Disk Mark as well as the bandwidth potentials of each pairing with the peak performance of 235 MB/sec coming from a solid-state USB 3.0 flash drive (Sandisk Extreme 64GB) to an SSD (Samsung 850PRO). Copying from the SSD to two different Seagate ST3000DM001 7200rpm 3TB HDDs gave similar results - irrespective of which SATA group (SATA or S_SATA), while copying from one ST3000DM001 to another (on separate SATA groups) yielded almost 180 MB/sec and when both ST3000DM001 disks were connected to the same controller (SATA), copy rate dropped a wee bit (~8% down) as could be expected due to drive contention on same controller.
 
So, the root cause of my earlier poor file copy performance had nothing to do with hardware from either Intel manufacture or Asrock implementation but would appear to have been due to a freeware copy utility I've used elsewhere with no similar challenges. All I can think of is that the version downloaded was corrupted, but I will - henceforth - avoid using this freeware copy tool and stick with the MS Windows File Copy.
 
To Parsec: My apologies for unfounded comments regarding "kludges"
To all here: My thanks for your inputs and tolerance - much appreciated
To unnamed dealer/neighbour: My sincere thanks for your time & effort - also much appreciated
 
All's well that ends well... Big smile
 
Dave


Edited by DevillEars - 01 Aug 2015 at 11:12pm
If music be the love of food, eat on
Back to Top
mikejonesny View Drop Down
Newbie
Newbie
Avatar

Joined: 25 Jan 2021
Status: Offline
Points: 1
Post Options Post Options   Thanks (0) Thanks(0)   Quote mikejonesny Quote  Post ReplyReply Direct Link To This Post Posted: 25 Jan 2021 at 1:34am
Does Asrock X99X has got support for SATA 3gbps hdd, I'm not a tech guy and just studying on the best budget options / alternatives for my current sys replacement, stumbled upon this post and thought I should better ask, im taking a look at various sata 3gbps options from allhdd here: https://www.allhdd.com/hard-disk-drive/sata-3gbps/ but do still in a decision mode, don't know whether I shoudl go for hp, seagate or dell 1tb hdds, can anyone confirm port availability for sata 3gbps i would be grateful, or if i dont get a reply ill post a new thread if needed.
be the best!
Back to Top
Xaltar View Drop Down
Moderator Group
Moderator Group
Avatar

Joined: 16 May 2015
Location: Europe
Status: Offline
Points: 22763
Post Options Post Options   Thanks (0) Thanks(0)   Quote Xaltar Quote  Post ReplyReply Direct Link To This Post Posted: 25 Jan 2021 at 5:21pm
Quote
Storage
- 10 x SATA3 6.0 Gb/s Connectors, support RAID (RAID 0, RAID 1, RAID 5, RAID 10 and Intel® Rapid Storage 13), NCQ, AHCI, Hot Plug and ASRock HDD Saver Technology
- 1 x eSATA Connector, supports NCQ, AHCI and Hot Plug
- 1 x Ultra M.2 Socket, supports M Key M.2 SATA3 6.0 Gb/s module and M.2 PCI Express module up to Gen3 x4 (32 Gb/s)

*SSATA3_3 connector is shared with the eSATA port, SSATA3_2 connector is shared with Ultra M.2 Socket

**RAID is supported on SATA3_0 ~ SATA3_5 ports only.


From the product spec sheet. So yes, it supports all SATA standards.
As for HDD brands, I would recommend Western Digital (WD), black, blue or green
depending on your budget and requirements. I have had nothing but bad luck with
Seagate for many years now.
Back to Top
 Post Reply Post Reply Page  <12
  Share Topic   

Forum Jump Forum Permissions View Drop Down

Forum Software by Web Wiz Forums® version 12.04
Copyright ©2001-2021 Web Wiz Ltd.

This page was generated in 0.145 seconds.